☆ Yσɠƚԋσʂ ☆@lemmygrad.ml to technology@hexbear.netEnglish · 6 days agoChatGPT's hallucination problem is getting worse according to OpenAI's own tests and nobody understands whywww.pcgamer.comexternal-linkmessage-square42linkfedilinkarrow-up184arrow-down10cross-posted to: technology@lemmy.ziptechnology@lemmygrad.mltechnology@lemmy.ml
arrow-up184arrow-down1external-linkChatGPT's hallucination problem is getting worse according to OpenAI's own tests and nobody understands whywww.pcgamer.com☆ Yσɠƚԋσʂ ☆@lemmygrad.ml to technology@hexbear.netEnglish · 6 days agomessage-square42linkfedilinkcross-posted to: technology@lemmy.ziptechnology@lemmygrad.mltechnology@lemmy.ml
minus-squareRedWizard [he/him, comrade/them]@hexbear.netlinkfedilinkEnglisharrow-up21·6 days agoYeah, they would have to have a sure fire way to identify the content as 100% AI generated so they could ditch it. If they’re training on Reddit Comments then they’re fucked.
minus-squareTankieTanuki [he/him]@hexbear.netlinkfedilinkEnglisharrow-up19·6 days agoTo be fair, they do thanks the gold for the stranger fuck CCP bacon.
Yeah, they would have to have a sure fire way to identify the content as 100% AI generated so they could ditch it. If they’re training on Reddit Comments then they’re fucked.
To be fair, they do thanks the gold for the stranger fuck CCP bacon.